The Psychiatric Rating Scale for Diagnostic Classification of Children and Adolescents Interrater Reliability

• Main Author: Henning, E. Glenn (Elbert Glenn)
• Other Authors: Lawlis, G. Frank
• Format: Others
• Language: English
• Published: North Texas State University 1984
• Subjects: DSM-III criteria; psychiatric rating scale; psychiatric diagnostics; child psychiatry; Child psychiatry > Classification.
• Online Access: https://digital.library.unt.edu/ark:/67531/metadc331462/

This study was designed to assess the reliability of "The Psychiatric Rating Scale for Diagnostic Classification of Children and Adolescents" as an instrument for determining diagnoses congruent with DSM-III criteria. In Phase I graduate students from a University doctoral program in psychology independently rated case vignettes and completed the 64-item rating scale to arrive at Axis I or II diagnoses consistent with DSM-III classifications for Disorders Usually First Evident in Infancy, Childhood, or Adolescence. Subsequent correlations to determine individual scale reliability yielded significantly positive correlations. Clinicians practicing in three diverse metropolitan mental health settings acted as raters in Phase II of the study. Paired raters jointly interviewed a total of 54 child or adolescent patients and independently completed the rating scale to arrive at Axis I or II diagnoses. These diagnoses were subsequently correlated with diagnoses previously obtained by traditional psychometric methods. Phase II interrater agreement was 92 percent for Axis I and II combined, with a .96 correlation. Rating scale diagnoses when correlated with traditional psychometic diagnoses yielded an overall rate of agreement on Axis I of 95 percent for Rater 1 and 90 percent for Rater 2 and correlations of .96 and .95 respectively. Clinicians were asked to rate a case vignette having previously been given an erroneous diagnosis. This attempt to assess rater expectancy effects yielded an agreement rate of 100 percent for the correct diagnosis. These results supported both major hypotheses of the study at a minimum of the .001 level of significance. They also confirmed a prior belief concerning limited intrusion upon rating scale reliability from rater expectancies. As a reliable and objective method of eliciting, structuring, and evaluating patient information, the rating scale could aide in reducing interdisciplinary interview variability and time expenditures among clinicians while providing an appropriate foundation for entry into treatment.